Better Search Engine for Zekr!
Hamid Zarrabi-Zadeh
Due to some limitations of the current "simple" search used in Zekr, we (Mohsen and I) have decided to replace it with a variation of the search engine used in Tanzil. Some of the features of Tanzil search engine are listed below:
- Simple and Powerful: despite its simplicity (with less than 50 lines of code), the search engine has powerful features like and/or/not operators, similar-letter match, wildcards, etc.
- Google-like Search: default behavior of Tanzil matches the ones used in Google. See for example the "Automatic And", "Exact Phrase" and "Not operator" features described here.
- Partial Diacritics: unlike Zekr's "match diacritics" option that requires users to enter all diacritics, Tanzil allows users to enter only a few diacritics to distinguish the words.
- Easy to Use: all features can be accessed through the search string itself, without the need to put an option menu or check-box. See for example how Exact Phrase and Whole Words can be defined in the search string.
- Efficient and Quick: the search engine responds to the queries with more than hundreds results in just a few milliseconds. Therefore, there is no need to put a limitation on the number of search results (like the current limit of 50 in Zekr).
- No Indexing Required: unlike Lucene search engine used in Zekr's advance search, Tanzil does not require any preprocessing time (and space) to create an index over the quran text (and separately for each translation text).
Regards,
--Hamid
Mohammad Derakhshani
On Thu, 2008-02-07 at 21:54 -0500, Hamid Zarrabi-Zadeh wrote:
> Hello All,
>
> Due to some limitations of the current "simple" search used in Zekr,
> we (Mohsen and I) have decided to replace it with a variation of the
> search engine used in Tanzil. Some of the features of Tanzil search
> engine are listed below:
> * Simple and Powerful: despite its simplicity (with less than 50
> lines of code), the search engine has powerful features like
> and/or/not operators, similar-letter match, wildcards, etc.
> * Google-like Search: default behavior of Tanzil matches the
> ones used in Google. See for example the "Automatic And",
> "Exact Phrase" and "Not operator" features described here.
> * Partial Diacritics: unlike Zekr's "match diacritics" option
> that requires users to enter all diacritics, Tanzil allows
> users to enter only a few diacritics to distinguish the words.
> * Easy to Use: all features can be accessed through the search
> string itself, without the need to put an option menu or
> check-box. See for example how Exact Phrase and Whole Words
> can be defined in the search string.
> * Efficient and Quick: the search engine responds to the queries
> with more than hundreds results in just a few milliseconds.
> Therefore, there is no need to put a limitation on the number
> of search results (like the current limit of 50 in Zekr).
> * No Indexing Required: unlike Lucene search engine used in
> Zekr's advance search, Tanzil does not require any
> preprocessing time (and space) to create an index over the
> quran text (and separately for each translation text).
> Please take a look at Tanzil search engine, and let us know if you
> have any comment or suggestion, before we port the implementation to
> Zekr.
I think the only feature we will miss is the fuzzy search provided by
the Lucene library.
I like that feature. Is it possible to have fuzzy search still?
Mohammad
Hamid Zarrabi-Zadeh
The plan is to keep the "advanced search" (Lucene engine) unchanged and add the new engine in place of the simple search. So, you can still use the fuzzy search feature in the advanced tab if you like.
--Hamid